🚀 go-pugleaf

Paul Carpenter <paul@pcserv.demon.co.uk> writes:

>chris@keris.demon.co.uk "Chris Croughton" writes:
>-No it doesn't with mail, because KA9Q sticks '>' on the front of any
>-line starting with "From ".  In news batches this doesn't happen because
>-it uses a different mechanism (the #! rnews line contains a character
>-count for the message so other #! rnews lines within it get ignored).

>Point 1 I was talking about mail, which KA9Q doe *NO* inserting of '>'
>when RECEIVING a message, you are refereing to REPLYING to a message within
>PCELM and the like!

If KA9Q SMTP server does not insert the > characters, then it is broken
(assuming that it holds the mailbox in "From " format).  The fact that your
mail reader may use the same character to indicate quoting is nothing to
do with it.  The SMTP server MUST guarantee that the body of the message
will not confuse the mail handling software into thinking it is a new
message.

The ONLY problem with this, is that ">From " and ">>From " etc. are not
escaped.  This means that the mail reader software cannot automatically
strip the byte stuffing.

If the SMTP server is not using "From " format, but the mail reader
requires it in From format, then the software that is performing the
conversion is broken.

>-There are lots of programs which hande mail fine with just the "From "
>-string (note, you have to match it at the start of the line, not
>-stripping off spaces or anything like that; use (Left$(lin$,5) = "From ")
>-to test, not Instr etc.) including PCELM, VIEW, John Washington's
>-mail2snz, my mail to news script, Tim Illingworth's Listserv software,
>-my list and mail server software, etc.

>I am fully aware of these items of software and I was referring to
>paragraghs in the _body_ of the text STARTING with 'From .....'

And these lines (in fact all lines starting with those magic 5 characters)
should've been escaped by the SMTP server.

>->The addendum I would make is that the next line contains a valid header code
>->to be a bit safer on counting the 'From' by counting 'From' plus next line has
>->started with 'RFC822 keyword:' then do the subtraction as before.

>Tell me how many of your message headers of RECEIVED emails do NOT start

>From xyz@isp.co.*......
>Received: ..........

All the ones in my mailbox.  My SMTP server stores the envelope to thus:

From somone@some.domain ....
Return-Path: <.....>
Received: ......


>-That *will* mess up, because several mailers include non-822 headers and
>-they may be before you find an 822 one.  You might be safe saying "at
>-least one RFC 822 header before you find a blank line" but it's really
>-not necessary.

>The non-RFC822 keywords are not _normally_ inserted en-route before the original

I would say that you cannot guarantee what order the headers will be in, or
even a partial order, except that the Received lines will be in hop order.


-- 
Stewart Brodie, stewart@wuzz.demon.co.uk
S.N.Brodie@ecs.soton.ac.uk  <-- prefer this one